Dr. Weng-Tink Chooi
Thu 11 Jan 2018, 13:10 - 14:00
S1 (7 George Square)

If you have a question about this talk, please contact: Anna Mas-casadesus (s1462664)

The cognitive training literature currently agrees that repeated practice or training on a task that increases in difficulty produces improvements only on the practiced/trained task. In other words, only near transfer effects are detectable. To date, no studies have empirically tested and explained the mechanism behind increased performance on a trained task, specifically the n-back task. The n-back task is a working memory task that engages the updating of information in one’s temporary online storage.  In this talk, I will present and discuss error distribution and changes in types of error made in a verbal n-back task using eight pairs of homophone words and incorporating position and word lures. The aim of study 1 was to investigate how improvements occurred by evaluating the error distribution for each load of the task (2-back, 3-back, etc.) over time. Participants in study 1 trained for 20 sessions (30 minutes/session) over five weeks. Results from 20 participants analyzed suggested that improved performance was due to reduced false alarms in both non-target and phonologically similar words accompanied by an increase in targets missed. We calculated d’ (hit rate – false alarm rate) and c, response bias, to track changes in participants’ response to the task and observed that improvements were due to a shift in responses that became more conservative as training progressed, hence the reduction in false alarms but also missing or rejecting more targets. In study 2, we included measures of simple span tasks (letter span and digit span) and an updating paradigm called Reference Back at pre- and post-training to evaluate if any changes or strategies developed due to training transferred to tasks that relied on serial order memory and/or updating.